Correlation of subjective sound perception of exhaust systems with sound quality metrics

In this contribution a method will be presented to assess the sound quality of automotive exhaust systems. The aim is to derive a measure (a model) that is able to predict the subjective perception of exhaust sounds. The approach is based on the use of sound quality metrics. The exhaust sounds are measured behind the cars (exterior noise) and recorded using an artificial head. Different tests are performed for several categories of cars and related exhaust systems. A subjective evaluation of these sound recordings is made by a jury using the pair comparison technique. From the same measurements a large set of metrics is computed. Using multiple regression analysis these metrics are then correlated with the subjective data (i.e., the results of the jury). This leads to a parametric model that can be used as a measure of the sound quality. The choice of the techniques used during the different steps will be motivated. Eventually, the further integration of sound quality in the development of an exhaust system will be discussed.